Treasures @ UT Dallas
Not a member yet
7697 research outputs found
Sort by
Remote Sensing and Spatial Distribution Patterns of Guyanese Resource Palms
No full textArecaceae (palms) are an essential resource for indigenous communities as well as fauna
populations in the Amazon. The present study conducts an extensive examination of the
application of remote sensing imagery for the detection of Arecaceae at both regional and
local scales within the previously understudied nation of Guyana. Subsequently, the detected
specimens of Arecaceae are used to answer critical questions concerning the distribution and
environmental drivers of their growth, and the impacts anthropological disturbance by way
of localized agriculture has on the regenerative abundance and diversity of Arecaceae
Photodynamic Modulation of the Tumor Stromal and Immune Microenvironment Using Engineered Nanoformulations
No full textDrug delivery in solid tumors is a considerable challenge. Desmoplasia, an excessive synthesis of
extracellular matrix, limits the delivery of drugs into tumors, restricts blood flow, increases the
solid stress of tumors, and leads to tumor progression. Photodynamic priming (PDP) is a
subtherapeutic iteration of photodynamic therapy that modulates the tumor vasculature, tumor
stroma, and immune microenvironment to improve the delivery of drugs and sensitize tumors to
secondary treatment modalities, such as chemotherapy, immunotherapy, and radiotherapy.
Visudyne, a liposomal formulation of Benzoporphyrin Derivative, is the only approved liposomal
formulation of a photosensitizer (PS). In this dissertation, we used PDP as a tool to manipulate the
tumor microenvironment in order to enable a number of cancer management strategies. These
include improving the time-to-surgery and diagnostic accuracy during fluorescence-guided
surgery using fluorescent molecular probes, improving the delivery of immune checkpoint
inhibitors while synergistically inducing immunogenic cell death, and enabling the self-delivery
of liposomes that carry immune checkpoint inhibitors to improve survival in pancreatic cancer.
This dissertation focused on the systematic engineering of liposomal-based PDP systems by
modulating photosensitizer composition and spatial localization and fine-tuning molecular
targeting of tumor receptors which can also serve as immune checkpoints. We also explored the
future potential for combining PDP using engineered liposomal systems with Onivyde (liposomal
irinotecan) as a means to synergically reduce cancer cell viability. Taken together, the results
presented in this dissertation outline implications for carefully engineering liposomal
nanoformulations as platforms for the modulation of the tumor microenvironment that lie far
beyond improving drug delivery and sensitization to secondary treatments. These implications
include transforming nanoformulations into tumor self-penetrating nanoformulations without the
need for additional photosensitizer administration and activation and neoadjuvant priming of
tumors to augment image-guided surgery; such applications have not been considered previously.
Overall, this dissertation paves a path for new and underexplored theranostic benefits of PDP using
nanoformulations, such as targeting and eliminating unresectable microscopic diseases, improving
the therapeutic efficacy of CAR T cells, protein-targeting chimeras (PROTAC), carbon ion beam
therapy, and the diagnostic accuracy of radionuclides, and radioimmunoconjugates
Giglio Albarelli: Cross-cultural Exchange and the Florentine Identity in Mamluk Ceramic Art
No full textThis thesis examines the historical and cultural significance of the Giglio albarelli, a group of six
fifteenth-century Mamluk ceramic vessels featuring the Florentine Giglio. Albarelli, originally
crafted as practical medicinal containers, evolved into symbols of cross-cultural exchange
between Renaissance Florence and the Mamluk Sultanate. Through an analysis of their design—
emphasizing the distinctive Giglio emblem—and the fusion of Chinese, Islamic, and European
artistic motifs, this study situates these objects at the intersection of trade, diplomacy, and artistic
innovation.
The research employs a multidisciplinary methodology, combining archival studies, visual
analysis, and first-hand examination of the albarelli housed in museums across Paris, Toronto,
Doha, and Abu Dhabi. By tracing their journey from functional apothecary jars to prestigious
diplomatic gifts or collectibles, the thesis highlights their dual nature as both utilitarian objects
and markers of cultural synthesis.
Key findings reveal that the Giglio albarelli may have served not only as practical vessels but
also as symbols of the flourishing relationship between Florence and the Islamic world. Their
standardized dimensions, shared heraldic motifs, and intricate decoration underscore the
craftsmanship of Mamluk artisans and their adaptation of Chinese-inspired designs to cater to
European tastes.
The Giglio albarelli embody the interconnectedness of the Mediterranean during the fifteenth
century, reflecting the dynamics of trade, cultural exchange, and mutual influence. By placing
these vessels within their broader historical context, this thesis contributes to a deeper
understanding of the role of Islamic art in shaping Renaissance material culture and cross-
cultural dialogues
A Novel Biomechanical Culture Platform to Explore the Structural Basis of Tensional Homeostasis in Soft Biological Tissues
No full textTissue homeostasis is an essential requirement for multicellular life and requires control over cell
proliferation, differentiation, paracrine signaling, and physical interactions with other cells and
with the surrounding extracellular matrix (ECM). Among the many physiological variables that
are under homeostatic control, the physical properties of most cells, tissues, and organs are
optimized to operate under a preferred state of mechanical tension. Much of what is known about
mechanical homeostasis of tissues under tension, or “tensional homeostasis,” is based on
experiments on tissue equivalents, e.g., fibroblast-populated collagen gels. Tissue equivalents
cultured in presence of a physical constraint at their boundary develop tensile forces under static
conditions, which has led to postulate the existence of a homeostatic tensile force. However, it is
conceivable that the homeostatic target is a material quantity, such as stress.
To gain a mechanistic understanding of how tensional homeostasis is developed and maintained
in healthy soft tissues, and subsequently disrupted by disease processes, there is a need for
improved tissue engineering platforms, also known as bioreactors. Recent studies have presented
bioreactors capable of culturing fibroblasts in 3D collagen matrices under both uniaxial and biaxial
tension. However, current bioreactors have limited imaging capabilities, with the acquisition of
structural information occurring only at discrete time points (due to the lack of sterility while
imaging). To address these limitations, we have developed the first miniaturized computer-
controlled biaxial bioreactor fully integrated with a confocal microscope. The bioreactor design
was inspired by devices in existing literature and built with the following objectives: (1) to allow
bioreactor functionality outside of a cell culture incubator; (2) to miniaturize the biomechanical
testing apparatus; and (3) to enable continuous confocal imaging functionality. Here, we present
the bioreactor design and showcase its ability to dynamically measure homeostatic force evolution,
structural remodeling of collagen, and cross-sectional area to estimate stresses in uniaxially
constrained tissue equivalents. In addition, we prove the feasibility of exploring the relationship
between structure and function in biaxially constrained tissue equivalents and discuss future plans
to enable biaxial stretching capability which will further expand the functionality of this new
platform to investigate fibroblast mechanobiology
Changing the Default of a Long, Violent History: Inclusive Language and Relational Coordination Theory
No full textIn this dissertation, I examine how organizations espouse intentions to value diversity and how
this intent often fails to translate into practices from these initiatives. As such, this research
examines the gap between normative beliefs (ones that are adopted across organizations) and
positive actions (empirically verified practices) which have resulted in conflicting discourse
regarding the benefits of diversity. Previous research tells us that, in the short term, the inclusion
of diverse backgrounds in the workforce is likely to hinder relational coordination as ingroup-
outgroup differences are perpetuated. Thus, actions to enlighten individuals about their relational
partner’s differences must be taken. I look toward identity disclosure as a mechanism which
allows individuals to inform others about their differences. Conceptually, I rely on language as a
mode of communication which opens the door to relational coordination. Empirically, I look
examine to relational coordination—a mutually reinforcing process of communicating and
relating for the purpose of task integration—as identity disclosure is inherently part of this
process. identity disclosure creates shared knowledge prior to system thinking by informing
individuals about each other’s identities. This information is critical to eventual communication
around tasks.
Relational coordination theory is critical to my dissertation because it begins with organizational
structures that bring together diverse workers and explains how they work together and relate,
eventually influencing work performance. In one qualitative study, a field study, and six
experimental studies, I explore diversity initiatives and the identity disclosure process.
Specifically, I examine the role of language in identity disclosure’s impact on relational
outcomes, and eventual performance. Taken together, these studies contribute to the relational
coordination literature by showing the effects of language and disclosure on organizational
relationships. These findings are critical in terms of building relationships across differences as
organizations try to leverage diversity within the workplace.
I begin by conceptualizing the influence of default changing vernacular and identity disclosure
within the organizational context and, I propose that the adoption of inclusive language in a
heterogeneous organizational environment can influence identity management. In turn, identity
disclosure will lead to greater societal change as these linguistic adoptions lead to changes in
default conceptualizations of stigmatized groups. I draw on and develop two theoretical
perspectives, relational coordination theory and stigma identity disclosure theory, to propose and
later test relational outcomes through language defaults. The qualitative data were collected from
97 LGBTQIA+ individuals and the experimental data was collected online through Prolific.
Next, I attempted to gain deeper insight into identity disclosure. Study 1 is a lab experiment
manipulating a supervisor’s use of inclusive language. Using a sample of 160 Lesbian and Gay
workers, I find that using the term significant other (an example of inclusive language) as
opposed to misgendering the subordinate’s spouse (e.g., referencing a gay man’s spouse as wife
instead of husband) leads to higher interpersonal awkwardness, lower trust, and increased
relational conflict. Study 2 extends these findings (n=368) by manipulating a 2 (Identity
Disclosure) x 2 (Inclusive Language) showing that interpersonal awkwardness and relational
conflict are highest when identities are not disclosed, and inclusive language are not used. Study
3 uses a Lesbian and Gay population sample (n=398) to test a 2 (Supervisor Identity
Receptiveness) x 2 (Inclusive Language) manipulation. I find similar results in terms of
interpersonal awkwardness and relational conflict but extend the prior findings to include
strategies for managing concealable stigmas. This experiment shows that subordinates are most
likely to assimilate—project the characteristics of a more socially valued group—when inclusive
language is used, and their supervisor is not receptive to their identity. This study highlights the
detriment of organizational adoption of normative diversity practices without positive actions to
change beliefs about outgroup individuals. Study 4 looks to a population sample of women
(n=283). This study uses a 2 (Supervisor Inclusive Language) x 2 (Supervisor Gratitude)
manipulation to examine supervisor-subordinate relationships based on the supervisor extending
gratitude for pregnancy disclosure. In this instance, I find relational conflict to be highest in the
no inclusive language/no gratitude condition and significant differences when collapsing across
gratitude. Study 5 looks to African American/Black employees (n=275). This study uses a 2
(Supervisor Inclusive Language) x 2 (Supervisor Receptiveness) manipulation to examine
supervisor-subordinate relationships based on the supervisor’s reaction to a parental leave policy.
Here I find that fear of disclosing the intent for parental leave is highest in the inclusive
language/no support condition. Finally, Study 6 looks to a population sample of gender non-
conforming employees (n=288). This study uses a 2 (Supervisor Identity Receptiveness) x 2
(Supervisor Learning) manipulation to examine supervisor-subordinate relationships based on
the supervisor’s attempt to learn about outgroup differences. As such, I find that relational
conflict is lowest in the receptive/learning condition. These findings show that positive actions to
learn about one’s identity who is different can impact a relationship even if one was not
originally accepting of the other’s identity. In the final chapter, I examine the full model using a
25 week field study. This study uses Discontinuous Growth Modeling (DGM) to measure
identity disclosure and the variables manipulated in the previous experiments across time. I
found that several of the experimental results held true. The implications of these findings and
future research directions are also discussed
Neural Mechanisms Underlying the Relationship Between Trait Mindfulness, Affect, and Cognitive Performance
No full textMindfulness is the ability of an individual to be aware of one’s internal and external experiences,
moment-by-moment, with a specific attitude of curious, non-judgmental acceptance. Trait
mindfulness refers to the dispositional ability of an individual to practice such awareness with a
specific attitude of nonjudgment and acceptance in daily activities. There has been a growing
recognition of the beneficial influence of mindfulness on mental health, both cognitive and
emotional. It is prescribed for treating and managing disorders in the ill, and improving
cognitive, academic and general well-being in the healthy. However, the mechanisms by which
mindfulness brings about its salutary effects are not yet understood. In pursuit of understanding
mindfulness mechanisms, the current studies examined the behavioral and neural processes
associated with trait mindfulness during performance of affective attention tasks.
Many psychological and neurocognitive mechanisms have been proposed to underlie the salutary
benefits of mindfulness. Two commonly proposed mindfulness mechanisms are complete
awareness to moment-by-moment experience, and acceptance. Importantly, behavioral evidence
suggests that, in high mindful individuals, there are changes in emotion regulation processes
across the time-course of experience that improve subsequent cognitive performance.
Specifically, I hypothesized that complete attention to present-moment events initially enhance
the emotional experience of an event affecting cognitive performance. This initial processing
affords subsequent regulation of the enhanced emotional experience through acceptance
strategies which then reduce the emotional influence allowing for better cognitive performance.
Results from two behavioral studies provide support for this hypothesis. I then examined the
neural bases of these processes using functional magnetic resonance imaging. Such emotional
experience and its subsequent regulation would be expected to be reflected in brain activity.
Emotional experience in relation to the self is characterized and measured by the extent of
activity in emotion generation regions such as the amygdala and insula. The ability to then accept
emotions as mere objects of mental experience and not as reflections or representations of the
self (i.e., a shift in perspective) may involve higher top-down processing mediated by the
prefrontal and cingulate cortices.
In the fMRI study, I examined mindful emotion regulation processes, operationalized as specific
neural changes occurring in emotion generation and regulation areas, across the duration of an
emotional Stroop task. Specifically, I hypothesized that trait mindfulness-related Blood Oxygen
Level Dependent (BOLD) signal in emotional generation areas would decrease across the
duration of the task, while BOLD signal in emotion regulation areas would increase.
Additionally, I hypothesized that such changes would predict both task-related and task-
unrelated (i.e., academic and verbal intelligence) performance. Results partially supported my
hypotheses and revealed complex time-course changes. Trait-mindfulness related increases and
decreases in BOLD signal were observed in both emotion processing and regulation areas,
specific to emotional demand. Trait-mindfulness related increases in Left Superior Frontal Gyrus
(LSFG) predicted Reaction Time (RT) improvements across the later task runs for negative
valence high arousal, compared to low arousal words. Further, trait mindfulness related
decreases in Left Superior Medial Gyrus (LSMG) during earlier task runs, in response to
negative low arousal compared to positive words, predicted higher academic performance (i.e.,
GPA).
In summary, results partially supported the hypotheses that mindful emotion regulation processes
comprise of reductions in emotion processing and increases in emotion regulation activity across
the duration of the task. Specific changes in these processes also predicted task-related and task-
unrelated performance, providing partial support to the idea that the neural changes indicating
mindful awareness and acceptance are related in specific ways to emotional experience and
behavioral performance
Building Robust AI Systems: Addressing Uncertainty, Data Noise and Scarcity in Modern Machine Learning
No full textThe increasing complexity of real-world machine learning applications, such as autonomous
systems, medical diagnostics, and natural language processing, demands models that can
operate reliably in environments characterized by uncertainty, data scarcity, and noisy or
ambiguous inputs. Traditional machine learning approaches, which rely on large, clean,
well-labeled datasets, often fail when faced with ambiguous inputs, limited labeled data, or
abundant but unlabeled data, leading to unreliable predictions and poor generalization.
This thesis addresses these critical challenges by developing robust learning frameworks that
enhance the uncertainty handling, adaptability, reliability, and performance of models in
such challenging environments. First, the Hyper-Evidential Neural Network (HENN) is
introduced to model vagueness uncertainty in classification tasks with composite class labels.
By leveraging Subjective Logic and Dirichlet distributions, HENN quantifies uncertainty and
improves decision-making in ambiguous data scenarios, such as medical diagnostics. Second,
NestedMAML, a nested bi-level optimization framework, is proposed to improve robustness
in corrupted few-shot learning with noisy and out-of-distribution tasks or instances. By
weighting tasks and instances during meta-training, NestedMAML reduces the influence of
noisy or irrelevant tasks and instances, improving robustness to distributional shifts and label
noise during meta-training, ensuring better generalization. Third, a semi-supervised meta-
learning framework, Platinum, is presented to leverage Submodular Mutual Information
(SMI) functions to select the most informative unlabeled data during meta-training in inner
and outer loops, ensuring that the model can leverage large amounts of unlabeled data while
minimizing the impact of noise, leading to better generalization in diverse tasks, even when
only a few labeled examples are available. These contributions provide a comprehensive
approach to tackle vagueness uncertainty, data scarcity, and noisy inputs in machine learning,
advancing methods for robust and adaptive learning in real-world applications
Medium Voltage Direct Current (MVDC) Power Cables for Wide-body All Electric Aircraft
No full textAll-electric aircraft (AEA) is a promising solution to achieve net-zero emissions transportation.
Among technical challenges in obtaining AEA, challenges related to power distribution and
protection system are of great importance, since it contributes to about 30% of the entire mass of
the electrical power system (EPS). Increasing the voltage level and developing bipolar MVDC
EPSs are solutions to decrease the total weight of the power distribution and protection system,
however, this solution exacerbates challenges in designing aircraft cable systems such as arc
tracking, partial discharges (PD), and thermal management. The study investigates the challenges
in designing bipolar MVDC cable systems for AEA and proposes multiple cable systems capable
of tackling those challenges while maintaining higher power density and lower system mass. To
develop the proposed cable system, first, the ampacity at reduced pressures of cruising altitude of
12.2 km is evaluated to the atmospheric one. Afterward, the effects of distance between the poles
in the bipolar cable systems are studied. Eventually, multilayer insulation systems are designed to
tackle challenges in designing aircraft cables while showing smaller mass and size compared to
existing ones
Development of Machine Learning Methods for Ultrasound RF Data Processing and Analysis
No full textThis thesis presents the development and application of machine learning methods to enhance the
processing and analysis of ultrasound radiofrequency (RF) data. The primary aim is to address
current limitations in ultrasound imaging by leveraging the rich information contained in RF data
and integrating it with advanced machine learning techniques. The research focuses on two main
areas: the detection and classification of novel ultrasound contrast agents and the tissue
differentiation in prostate imaging. Novel contrast agents, such as chemically crosslinked
microbubble clusters (CCMCs), were synthesized and characterized. These agents exhibit unique
acoustic signatures that, when analyzed using machine learning models like one-dimensional
convolutional neural networks (1D CNNs) and anomaly detection models could be distinguished
and separated. These unique acoustics when compared with the individual contrast agents,
exhibited higher energy which can potentially lead to improved contrast of the image and
visualizing the microvascular structures and tissue perfusion by using these contrast agents for
super-resolution ultrasound imaging. In addition to these novel contrast agents, we explored
hemoglobin microbubbles (HbMBs) as oxygen sensors to detect oxygen levels in the surrounding
by using the acoustic response of HbMBs in varying oxygen levels environment by using pixel
intensity comparison and 1D CNN model. We also explored the potential of differentiating
prostate peripheral zones and stromal regions using ex-vivo prostate RF data. Machine learning
algorithms, combined with radiomics features extracted from RF data, were employed to
accurately differentiate between various tissue types. The findings of this research underscore the
potential of integrating RF data with machine learning models to overcome the inherent limitations
of conventional ultrasound imaging. By enhancing image quality and diagnostic accuracy, these
advancements pave the way for more effective, non-invasive diagnostic tools in clinical practice,
ultimately improving patient outcomes and healthcare efficiency
Product Superposition With Transmitter State Information and Its Application in Reconfigurable Intelligent Surfaces
No full textLinks in wireless networks may exhibit non-identical coherence times, bandwidths, or spatial
correlations when nodes experience different scattering environments or have varying mobility.
This can happen in MIMO (multiple-input multiple-output) communications, with or without
reconfigurable intelligent surfaces (RIS), and in high-mobility scenarios. This dissertation
investigates the multi-user downlink MIMO channel under conditions where links have non-
identical coherence times and bandwidths, focusing on scenarios with channel state feedback,
as well as RIS-assisted communications under non-identical coherence times and unequal
correlation conditions. The study finds that exploiting the disparities in these diverse fading
conditions across multiple users can yield significant gains compared to techniques that do
not take advantage of these disparities.
For multi-user downlink MIMO under unequal coherence times, coherence bandwidths, and
channel state feedback, we propose a novel and efficient pilot-domain non-orthogonal multiple
access (NOMA) signaling scheme. Channel state feedback is considered in both perfect and
imperfect forms. Our method enables the combination of gains from product superposition
with those from transmit beamforming, which was not feasible with previous approaches. The
technical innovation of this work lies in reconciling the requirements of product superposition
and beamforming through new methods, allowing for the simultaneous harvesting of both
classes of gains. We demonstrate the advantages of the proposed approach in terms of
achievable degrees of freedom (DoF) and rates.
For downlink RIS-assisted communications, where the training overhead is usually higher
than in typical MIMO systems due to the large number of unknown channels induced by the
RIS, we propose a novel pilot-domain NOMA technique that can reduce this overhead. This
approach is particularly effective under conditions of unequal coherence times. By jointly
optimizing the transmit beamforming and RIS reflection coefficient vectors, our method
maximizes the achieved sum-rate, delivering gains in both rate and degrees of freedom. The
proposed scheme includes efficient pilot placement strategies for multi-user RIS-assisted
systems with varying coherence intervals.
In the context of downlink RIS-assisted communications, we explore the channel training and
estimation gains that arise from the disparity in antenna correlations, known as correlation
diversity. For a frequency-division duplex (FDD) system, we conduct a comprehensive study of
channel state estimation, feedback, and joint beamforming in the presence of RIS and antenna
correlations. Our pilot configurations are optimized based on the degrees of freedom, and
sum-rates are optimized through beamforming. The proposed approach identifies and analyzes
the common and disjoint eigenspaces of the correlation matrices, utilizing the technique of
product superposition. We first elucidate the fundamental aspects of the problem by deriving
achievable degrees of freedom and rate regions for two users, and then extend these results to
an L-user RIS-assisted downlink MIMO setting, a natural scenario for correlation diversity